> ## Documentation Index
> Fetch the complete documentation index at: https://docs.kime.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Agentic Workforce

> See the work KIME's AI agents do for you: recurring analysis, agent runs and finished deliverables.

Agentic Workforce, also called **AWX (Agentic Workforce Execution)**, is the part of KIME that does the work for you. KIME's agents collect and analyze your data on a schedule, and turn [Actions](/features/actions) into finished deliverables such as articles and outreach emails. The Agentic Workforce page is where you follow all of it.

You find it in the sidebar under **Agentic Workforce Execution > Agentic Workforce**.

<Note>
  Agentic Workforce is in beta. Running a task uses **credits**. Your remaining credits are shown at the top of the page.
</Note>

## What you see

### Latest completed deliverables

The most recent outputs your agents have finished, with their **Format** (for example Article or Email) and **Completed** date. Click one to open it in [Deliverables](/features/deliverables).

### Recurring tasks

A weekly calendar of the work KIME runs for you automatically, for example:

* **Collect prompt data**: running your tracked prompts across the answer engines, every day.
* **Analyze AI Perception**: updating how AI describes your brand.
* **Reevaluate actions**: reviewing your data and suggesting new or updated [Actions](/features/actions).
* **Research new prompts**: finding new prompts worth tracking for [Prompt Explorer](/getting-started/prompt-explorer).

### Agent runs

Every deliverable an agent is working on or has finished. In-progress runs appear first, then recent completed and failed runs. Each run shows the steps the agent goes through. For an article these are:

<Steps>
  <Step title="Investigate search intent">
    Works out what people asking the prompt actually want to know.
  </Step>

  <Step title="Build platform strategy">
    Decides how the content should be shaped to be picked up by AI answer engines.
  </Step>

  <Step title="Gather sources">
    Collects facts and references to support the content.
  </Step>

  <Step title="Build outline">
    Structures the article.
  </Step>

  <Step title="Draft article">
    Writes the first draft in your brand's tone of voice.
  </Step>

  <Step title="Fact-check and revise">
    Checks claims against the gathered sources and fixes errors.
  </Step>

  <Step title="Humanize">
    Polishes the language so it reads naturally.
  </Step>
</Steps>

### Shortcuts

* **Content** opens your content settings, such as content language, content profile and email senders.
* **Learnings** opens the rules your agents follow when they write. See [Workspace settings](/administration/workspace-settings#learnings).

## How to set it up

<Steps>
  <Step title="Check your content profile">
    In **Workspace settings > Content**, review the **Content profile** KIME built from your website: tone of voice, writing style, audience and vocabulary. Edit it if anything is off, because every draft is based on it.
  </Step>

  <Step title="Set the content language">
    Choose the default language for generated articles in **Workspace settings > Content**.
  </Step>

  <Step title="Add email senders">
    Add the name and role that should sign outreach emails.
  </Step>

  <Step title="Add learnings">
    Add any rules you already know, for example terms to use or avoid, or facts that should always be mentioned.
  </Step>
</Steps>

## How to use it

<Steps>
  <Step title="Pick tasks in Actions">
    Go to [Actions](/features/actions), accept the suggested tasks you want and click **Start** on a task to let an agent do it. Each run uses a credit.
  </Step>

  <Step title="Follow the run">
    Watch progress under **Agent runs** on the Agentic Workforce page.
  </Step>

  <Step title="Review the result">
    When the run is done, the task moves to **In Review**. Open the deliverable, refine it and export it. See [Deliverables](/features/deliverables).
  </Step>
</Steps>

## Get the most value

<AccordionGroup>
  <Accordion title="Invest in your content profile and learnings" icon="graduation-cap">
    The better your content profile and learnings, the less editing each draft needs. Every time you correct something, save it as a learning so the agents get it right next time.
  </Accordion>

  <Accordion title="Run several tasks at once" icon="layer-group">
    Agents work in parallel, so start a batch of related tasks together, for example a set of buyer's guides for one product category.
  </Accordion>

  <Accordion title="Plan your credits" icon="coins">
    Credits are shared across your organization and refill monthly. Prioritize the tasks with the biggest expected impact on visibility.
  </Accordion>

  <Accordion title="Check failed runs" icon="triangle-exclamation">
    Failed runs are listed under Agent runs. If a run fails, check the task in Actions, or contact support if it keeps happening.
  </Accordion>
</AccordionGroup>

## Who it is for

* **Content and SEO teams** that want more content produced without adding headcount.
* **PR and outreach teams** that want ready-to-send pitches to editors and publishers.
* **Agencies** delivering AI visibility work for many clients.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.